2. Pan-PI3K Inhibitors
Activating mutations in PIK3CA, the gene encoding for the p110α catalytic subunits of PI3K, are associated with the growth and survival of cancer cells that play a role in cell survival, proliferation, differentiation, and glucose transport
[4][10][11][12]. Dysregulation of the PI3K-pathway may also contribute to resistance to a variety of anticancer agents
[13][14]. First-generation PI3K inhibitors (PI3Ki), also known as pan-inhibitors of PI3K, target all four catalytic isoforms of class I PI3Ks (α, β, γ, and δ)
[15][16]. These inhibitors encompass a broad spectrum of activities, and this broader inhibition leads to a higher risk of adverse events (AEs) and off-target toxicity, which have limited their use at therapeutic doses and caused treatment discontinuation
[17][18][19].
2.1. Pan-PI3K Inhibitors in HR+/HER2− Breast Cancer Subtypes
In luminal BC subtype PIK3CA activating mutations occur in roughly 40% of cases and, in this setting, the pan-PI3Ki mostly include buparlisib (BKM120) and pictilisib (GDC-094).
2.1.1. Buparlisib
Buparlisib (BKM120) is an orally selective pan-PI3Ki that inhibits all four class I PI3K isoforms (p110α, β, γ, and δ), as well as somatically mutated p110α, with no activity against the class III PI3K, or mammalian target of rapamycin (mTOR). In cellular assays and in vivo models of human cancers, buparlisib showed significant dose-dependent tumor growth inhibition, potent antiproliferative and proapoptotic activities
[20][21]. Miller et al. reported that the hyperactivation of the PI3K pathway, or PTEN loss expression, promoted antioestrogen resistance in HR+/HER2− MBC, inducing ER-transcriptional independent activity and growth of BC cells
[22][23]. Therefore, buparlisib plus fulvestrant has been successfully evaluated in a phase I study in postmenopausal women with HR+ MBC, previously treated with endocrine therapy
[24]. PIK3CA mutations proved to be a biomarker of poor prognosis, such as PgR negativity, TP53 mutations, and loss of PTEN expression. Conversely, mutations in AKT1 and ESR1 did not prevent tumor response. Treatment-related adverse events (AEs) were generally mild and included hepatic toxicity, rash and hyperglycemia
[25][26].
The phase III BELLE-2 clinical trial investigated the efficacy and safety of buparlisib plus fulvestrant in HR+/HER2− post-menopausal MBC patients, who were progressive during/after treatment with aromatase inhibitors (AI) and had received up to one prior line of chemotherapy
[27]. Patients were randomized to receive fulvestrant plus buparlisib or placebo and stratified according to PI3K status in tumor tissue by Sanger sequencing, and visceral disease status. The trial met its primary endpoint demonstrating a relatively modest benefit in progression-free survival (PFS) (6.9 versus 5.0 months compared to placebo; HR 0.78 CI 0.67–0.89;
p < 0.001). Many patients, in the buparlisib arm, discontinued the study early (1.9 months) due to adverse events (higher transaminase levels, hyperglycemia, rash, and mood disturbances) probably influencing the potential benefits of combination therapy. In the exploratory analysis on PIK3CA mutations identified in circulating tumor DNA (ctDNA), the combination treatment was associated with an improvement in PFS compared with fulvestrant alone only in patients with PIK3CA mutations. In contrast, prespecified analyses of PIK3CA mutation and/or loss of PTEN expression by Sanger sequencing in archival tissue samples did not show increased PFS with buparlisib, potentially suggesting a tumor evolution from initial diagnosis and treatment beginning.
Likewise, BELLE-3 trial explored the ability of fulvestrant plus buparlisib to restore endocrine sensitivity, compared to fulvestrant alone, in patients with luminal MBC subtype previously treated with ET and mTOR inhibitors
[28]. The trial showed that the combination therapy was associated with better PFS (median 3.9 months versus 1.8 months; HR 0.67 CI 0.53–0.84;
p < 0.001), particularly in patients with real-time polymerase chain reaction (RT-PCR) or ctDNA PIK3CA mutations. In ctDNA-detected PI3K alterations the median PFS was 4.2 months versus 1.6 months (HR 0.46 CI 0.29–0.73;
p = 0.00031) while it was 4.7 months versus 1.4 months (HR 0.39 CI 0.23–0.65;
p < 0.001) in the group tested by RT-PCR. On the contrary, in the BELLE-2 trial, patients with PIK3CA mutations tested in tumor tissue, PFS was not different between treatments. The discrepancy between the two studies could be explained by the different sensitivity of the test used for PIK3CA assessment, or by the absence of stratification, based on ctDNA at randomization and subsequently added as a secondary endpoint.
Additionally, Baselga et al., in the combined analysis from BELLE-2 and BELLE-3 trials, observed that assessing PIK3CA-mutational status in ctDNA resulted in larger clinical benefits compared with tissue samples PCR
[29][30]. The safety profile of buparlisib was consistent in BELLE-3 and BELLE-2. Psychiatric complications, related to the ability of buparlisib to penetrate the blood-brain barrier
[31], such as depression, anxiety, and suicidal ideation, were reported in 2% of patients enrolled in the BELLE-3 trial, in contrast with the absence of suicidal ideation in BELLE-2 study.
In phase II/III clinical trial BELLE-4
[32], the combination of paclitaxel plus buparlisib in MBC as first-line treatment was tested, showing that the addition of buparlisib to paclitaxel did not improve median PFS when compared to paclitaxel alone (8.0 months versus 9.2 months in buparlisib and placebo group, respectively; HR 1.18 CI 0.82–1.68). Furthermore, the HRs, in the group of patients with PI3K mutations, PTEN loss expression and PI3K wild type, were similar to the overall enrolled population. Buparlisib treatment was associated with a higher frequency of serious AEs (30.2% in the buparlisib arm versus 20.9% in the placebo group) and more than 40% of patients, in the combination arm, experienced diarrhea (55%), rash (43%), hyperglycemia, and nausea (41%).
2.1.2. Pictilisib
Pictilisib (GDC-094) is an oral class I PI3K pan-inhibitor. In vitro experiments showed preliminary antitumor activity (equipotent inhibition of the p110α and -δPI3K isoforms and less potent inhibition of p110β and -γ isoforms) and, in human tumor xenograft murine models, pictilisib demonstrated a strong inhibitory effect on the growth of human U87MG glioblastoma and IGROV1 ovarian cancer
[33].
In the dose-escalation phase I clinical trial, Sarker et al.
[34] evaluated the preliminary clinical activity of pictilisib in sixty unselected and heavily pretreated patients with advanced solid tumors. A low response rate, consisting of one partial response (PR) and one stable disease (SD), was observed. The toxicity profile of pictilisib was similar to buparlisib and the most frequent grade 3/4 AEs were rash, hyperglycemia, and pneumonitis. Pictilisib-related hyperglycemia was limited to grade 1–2; grade > 3 hyperglycemia occurred only in one patient. Mood alterations were not significant due to the low central nervous system (CNS) penetration of pictilisib compared to Buparlisib.
According to previous results by Schmid et al., in a phase II, open-label, randomized trial
[35], the addiction of pictilisib to anastrozole determined a synergic effect and was associated with increased inhibition of tumor cell proliferation over anastrozole alone in postmenopausal women with luminal-B early BC.
In the two-part, randomized, double-blind, placebo-controlled phase II clinical trial FERGI,
[36] investigating the benefit of adding pictilisib to fulvestrant in postmenopausal patients with HR+/HER2− MBC resistant to AI, no difference in median PFS was found in the combination treatment arm, regardless of PI3KCA mutation status. The median PFS was 6.6 months and 5.1 months in the experimental arm and control arm, respectively (HR 0.74 CI 0.52–1.06;
p = 0.096) in ITT-population and 6.5 months and 5.1 months, respectively (HR 0.73 95% CI 0.42–1.28;
p = 0.268) in women with PIK3CA-mutations. Pictilisib treatment was affected by drug toxicity, potentially limiting its efficacy. Rash, diarrhea, increase transaminase level, hyperglycemia, pneumonitis, bronchopneumonia, pleural effusions, and fatigue were the most common AEs in the pictilisib arm and in approximately 60% of cases grade 3 or worse adverse events were observed, with serious AEs occurring in up to 16% of participants.
Another phase II randomized placebo-controlled clinical trial, PEGGY
[37], failed to meet the primary endpoint. This, similar to BELLE 4, was conducted to evaluate the benefit of adding PI3K-inhibitors to paclitaxel vs. paclitaxel alone in pre-and post-menopausal HR-positive/HER2-negative advanced BC patients. Genetic alterations in the PIK3CA gene were assessed in about 35% and 32% of patients in both arms, respectively. At the interim analyses, no significant differences in terms of PFS and overall response (ORR) were observed, irrespective of PIK3CA mutation status. Indeed, the addition of pictilisib to paclitaxel did not improve median PFS in the entire (8.2 months vs. 7.8 months; HR 0.95 CI 0.62–1.46;
p = 0.83) and in the PI3K pathway-activated study population (7.3 months vs. 5.8 months; HR 1.06 CI 0.52–2.12;
p = 0.88). Toxicity due to pictilisib administration included maculopapular rash, hypertension, and hyperglycemia, and experimental treatment was also associated with higher grade ≥ 3 AEs, dose reduction, and discontinuation.
2.2. Pan-PI3K Inhibitors in HER2+ Breast Cancer Subtypes
Although the anti-HER2 agents demonstrated strong efficacy in patients with HER2 + BC, de novo or acquired resistance to HER2-target therapies represents a wide field of research. About 25% of HER2 + BC harbor PIK3CA mutations confer resistance to anti-HER2 therapy and poorer prognosis
[38][39][40][41][42][43]. Preclinical studies showed that HER2-signalling is extremely deregulated and is largely mediated by p110α rather than one of the other class-I PI3K isoforms, providing a robust rationale to target the PI3K pathway
[44][45][46]. The tolerability and activity of buparlisib in combination with trastuzumab were investigated in a phase Ib/II dose-escalation trial in PIK3CA unselected HER2-positive BC, progressive to trastuzumab. The recommended dose of buparlisib was reached at 100 mg/day plus weekly intravenous trastuzumab but, even if evidence of clinical activity was observed, the study did not meet the pre-established primary endpoint such as objective response > 25%
[47][48].
Furthermore, Guerin et al. studied buparlisib in combination with lapatinib, the orally dual anti-HER2/HER1 tyrosine kinase inhibitor, in the phase Ib/II trial PIKHER2, in HER2-positive trastuzumab-resistant MBC patients, independently to PIK3CA mutational status
[49]. Twenty-four patients were treated at five different dose regimens; the selected dose was 80 mg/daily for buparlisib and 1000 mg/daily for lapatinib. Main drug-related adverse events leading to discontinuation of treatment were gastrointestinal disorders, skin rash, depression and anxiety. The disease control rate was 79% (95% CI 57–92%) including 4% of CR and a 29% of clinical benefit rate (CBR) (95% CI 12–51%).
2.3. Pan-PI3K Inhibitors in Triple Negative Breast Cancer Subtypes
TNBC represents the most aggressive BC with an incidence of 15–20% and a high heterogeneity in the mutational profile. Due to limited possibility of target therapies, standard chemotherapy still represents the milestone of treatment
[50]. Approximately 10% of TNBC harbors a germline mutation in BRCA1 or BRCA2 genes and recent evidence has shown that platinum-based therapy offers promising activity both in early and metastatic settings
[51][52][53][54][55][56][57]. In TNBC, activating PIK3CA mutations, the majority located in the p110α subunit, are the second most frequent molecular aberrations after TP53 mutations and occur in 7–9% of primary TNBC, with a likely higher rate in advanced TNBC
[1]. PIK3CA mutations, with additional inactivating alterations in PTEN
[58] and activating mutations in AKT1, globally occur in about 25% of TNBC
[59]. In this subtype, dysregulation of the PI3K pathway has been correlated with chemotherapy resistance
[60] and loss of PTEN function confers resistance to PDL1-blockade and leads to increased PI3Kβ signaling. Indeed, based on preclinical models, combined therapy with an anti-PDL1 agent and a PI3Kβ inhibitor showed an improvement in cancer growth inhibition
[59][61]. However, the role of the PI3K pathway-targeted therapy in TNBC is also unclear. The activation of the PI3K pathway is more linked with the androgen receptor-positive subtype of TNBC and less correlated to TNBC compared to HR-positive and HER2-positive BC
[59][62]. In the BELLE-4 trial TNBC patients (about 25% of total) tended to have a worse prognosis with buparlisib versus placebo arm (5.5 versus 9.3 months; HR 1.86 CI 0.91–3.79) and versus HR+ population (9.2 vs. 9.2 months; HR 1.00 CI 0.66–1.529). In this trial, the poorer prognosis in the TNBC subgroup, may be explained by a lower duration of paclitaxel exposure in the buparlisib group suggesting that the toxicity of buparlisib may have compromised the adequate administration of chemotherapy; hence, researchers failed to confirm the PIK3CA mutation’s predictive role in TNBC subtype
[32]. In preclinical models, the pan-PI3Ki BKM120 sensitized BRCA-proficient TNBC to PARP inhibitor olaparib: researchers demonstrated that the dual PI3K and PARP inhibition significantly downregulated BRCA1/2 expression and reduced tumor cell growth
[63]. Likewise, a second study showed synergic activity of buparlisib combined with olaparib in an MMTV-CreBRCA1f/fp53 +/− mouse model of BC
[64]. Additionally, Matulonis UA et al., in a phase I dose-escalation trial, demonstrated anticancer activity in BC (54% of which TNBC) and ovarian cancer and in both germline BRCA (gBRCA)-mutated and gBRCA-wild-type patients
[65].